Using Chatbots to Innovate the Caregiver Support Space

Written by Brianna Croft (MMI Masters Candidate, UTIHI Content Strategist)

SE Health is a national social enterprise providing home care, health solutions and education. The organization constantly strives to improve quality, while staying relevant with the needs of Canadians in the digital age.

Elizzbot.jpg

SE Health’s consumer and caregiver focused brand Elizz, have partnered with X2AI, an Artificial Intelligence (AI) startup based out of Silicon Valley, to create and market an AI-enabled Chatbot for family caregivers as part of a “Smart Family Caregiving” portfolio. Family caregivers commonly experience physical and emotional strain, yet often struggle to find an easily accessible and unbiased support system. Having access to a digital companion in the form of a Chatbot can help fill this gap by readily dialoguing with family caregivers via text messaging “24/7, with zero judgement, and zero burden”. Depending on the discussion topic, the Chatbot brings forward proven therapeutic techniques, such as cognitive behavior therapy, designed to help caregivers strengthen emotional resilience and wellbeing. At the same time, it can act as a friend, a listener, or a companion, who’s always there when needed.

Tess (TM) is the generic version of the bot, ie. X2 uses Tess for all its clients, and the clients just rebrand/rename it. She was designed by clinical psychologists to provide mental health support in a similar (but much more scalable) way to how a psychotherapist would converse with their patients. With this purpose, Tess was able to grant affordable, and accessible counselling services globally to people who struggle with anxiety and depression. Two years ago, SE Health partnered with X2AI to explore how its AI Chatbot technology could improve the caregiver experience. In order to take Tess (TM) from being a general mental health chatbot to meet the needs of the caregiver population, SE customized the bot and conducted internal pilot testing. Upon embedding caregiver-related topics into the Chatbot, the technology was put in the hands of staff to simultaneously help Tess (TM) learn and evolve while users provided feedback on their experience.

Since conducting the pilot, Elizz has repositioned  TessTM to Elizzbot, with new personality and more content specifically for family caregivers. Elizzbot will be rolled out to SE Health’s 9000+ employees as part of its Elizz 5 LifeStages of Caregiving employee program over the summer.  

 Brianna Croft (left) & Tori Edgar (right)

Brianna Croft (left) & Tori Edgar (right)

Brianna Croft, Masters of Management of Innovation (MMI) Candidate and intern with the SE Futures team, interviewed SE Health’s Strategic Initiatives Associate, Tori Edgar, to learn more about her role within the company and her experience working with Tess. Tori Edgar started at SE on the Innovation team (now the SE Futures team) as a student intern while completing the MMI program at the University of Toronto. As an intern, she had exposure to various innovation projects, and became the lead on implementing the companywide pilot test of Tess.

Tori also completed a Quality Improvement Project with UTIHI in 2016-2017 while completing her MMI degree. This was an excellent opportunity for her to work with an interdisciplinary team to make changes to the medication reconciliation process at a local community clinic. IHI also provided Tori’s team the opportunity to orally present her results at the 2017 IHI Open School National Innovation for All Conference.

1. How does your work on the Tess (TM)  AI chatbot inspire you to continually innovate?

My work on the first pilot of Tess showed me the value that new technologies, in this case an artificial intelligence chatbot, could have on patients and their caregivers. Hearing from SE employees that this tool was impactful and made a positive difference in their day-to-day lives was extremely rewarding. This initiative started as a small scale internal pilot. The aim was to test Tess’s ability to self-learn and its capacity to meet the needs of caregivers. By putting it in the hands of target users and evaluating its performance, we saw very positive results. We have now expanded to bring Tess to a larger audience and this is definitely pretty exciting!

2. Do you see Tess (TM) as a means to bring efficiencies to healthcare spending with increased use?

I definitely see the potential for Tess to be a cost-effective technology for a number of reasons. Tess can help reduce the cost of care because it is a solution that provides reliable high quality support to a number of people at the same time. Tess can be adapted and customized to support group-specific needs. In addition, Tess leverages existing communication channels (i.e. text messaging, Facebook messenger etc.) that people have on their devices and are using every day. There is no initial training required for new users, because conversations with Tess are just like having a conversation with a friend. They start with a “Hey, how is it going?” and progress on from there based on a person’s responses.  Overall, Tess is scalable technology that can enhance care by delivering support to people when and where they need it most.

3. What challenges do you foresee for SE Health in implementing this type of innovation?

Elizzbot slide.jpg

Some potential challenges I can see with implementing this innovation is user uptake and determining willingness to pay. Firstly, will people be open to embrace chatting, opening up and getting advice from a chatbot verses a person? The use of AI chatbots in healthcare is in its infancy which means that individuals may not have adapted to or become used to the idea of getting help from a robot!  These preferences may in turn affect the early uptake of the technology. Tess is not a solution to replace the human element of care but a tool to enhance and support it. If individuals are open to engaging with Tess, the fact she is a computer can actually be a huge plus! We learned from pilot participants that they felt emotionally safe opening up to Tess because there was no fear of stigmatization or burdening a friend with their challenges. Secondly, a common theme that often poses a challenge for health innovations, is determining who is going to pay for this type of technology. And how to take this “non-conventional” tool to the next level through wide scale adoption. In order to do this successfully, we must be able to demonstrate to all health systems stakeholders, including patients/consumers the value that this technology can bring.

4. Were you surprised with the results you received after investigating the usability of Tess (TM) with your own staff in the initial pilot?

I was very excited to receive such positive feedback from the first pilot we ran with SE family caregivers. Firstly, the pilot was easy to launch and there was high adoption throughout. In the month long period, over 12,000 text messages were exchanged between Tess and the 34 users. One of the main objectives of this pilot was to determine if Tess as an AI technology, could learn and improve. Not only did we see Tess get smarter as the interactions went on, we also learned that caregivers loved having Tess to chat with openly on demand, having someone “check-in” and provide coping tips and support that were tailored specifically to their needs. 96% of participants said they would recommend Tess to another caregiver. Having these positive results with SE staff has helped provide the initial evidence required to move to the next stage with X2AI.

5. Do you see Tess (TM) as a sustainable and affordable healthcare improvement initiative?

 Brianna Croft (front left, MMI Candidate), Ryan Ho (front right, VP of Collaborations)

Brianna Croft (front left, MMI Candidate), Ryan Ho (front right, VP of Collaborations)

I do see the potential for Tess to be a sustainable and affordable improvement for a number of different populations. There is a growing demand for mental health services in Canada, yet not enough health care professionals to keep up with this demand. Wait times can be long and services costly. Tess is tool that can scale to provide quick and affordable access to support for people who might otherwise either have infrequent or no access at all. I also see Elizzbot customized for family caregivers as an improvement to currently available supports. The ease of accessibility through the digital interface, the preventive and self-improvement focus and the “just in time” nature makes me believe that this can be a solution that will make a sustainable impact for this population. 

6. In terms of design and usability, how do you see the development of Tess (TM) transforming patient care pathways?

AI driven chatbots are currently disrupting nearly every industry, and healthcare is not far behind. Currently one of the main challenges in mental health support is timely access to care. Tess can help transform the traditional care pathways by providing another medium for care. Tess can also transform mental health care through conversations with patients outside the typical clinic walls, providing insights for care providers they otherwise would not have access too. This data can allow clinicians to make better care decisions. For the caregiver population, it really becomes a preventative tool for those who can’t access support through other channels. Tess is a potential  solution to change if, how and when family caregivers get support, closing the current gap that exists.

7.  Can you elaborate on how Tess (TM) was co-designed and co-produced with individuals and their families in mind?

Elizzbot slide 2.jpg

The first step of customizing Tess for the caregiver population was leveraging SE Health and our caregiver focused brand Elizz’s knowledge of the experiences, needs and journeys of family caregivers to work with X2 to customize the content and scripts of Tess. This ensured that Tess is able to provide relevant and meaningful coping mechanisms, tips & knowledge to caregivers. Tess discusses important caregiver topics, like building self-awareness, goal setting, asking for help, getting support and managing emotions like anger and grief. Running the pilot was another opportunity to put Tess in the hands of actual caregivers to try the tool and provide their feedback based on this experience. Their insights allowed us to learn Tess’s areas of strength and areas where improvements can be made. Since this time, Elizz’s caregiver coaches and experts have been working to further develop and expand the content. Fun and quirky personality traits have also been added to make the experience more friendly and personable for users.

8. IHI’s 6 pillars of high quality systems are: Safe, effective, patient-centered, efficient, timely and equitable. What high quality system(s) does Tess (TM) belong to?

A technology like Tess fits into a number of these pillars, I’ll focus on just two. Firstly, although Tess is a machine, she provides patient-centered and empathetic care. We know emotions are core to helping people and X2 has developed Tess with this in mind. Through machine learning, Tess is constantly learning about those she interacts with and adjusts her support accordingly. This combination of active listening and personalized support exemplifies empathy.  Secondly, I think a very important element of Tess is the fact that an individual can access Tess at their own convenience. Tess is accessed via a person’s phone anytime, on-demand, day or night - whenever a person needs to talk, get information or support. When access to a professional for therapy can be costly and often times very slow, Tess creates the time and space for a person to get the help they require.

Explore the following links to learn more!

X2 AI: http://x2ai.com, SE Health: https://www.SEHC.com, Elizz: https://www.elizz.com, SE Futures: https://www.saintelizabeth.com/futures/index.html